It is currently Sun Nov 23, 2014 3:59 pm

All times are UTC - 4 hours [ DST ]





Post new topic Reply to topic  [ 12 posts ] 
Author Message
 Post subject: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 12:08 pm 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
and here we go gentlemen: the site that gives the football nerds boners, the all informed

football outhousers.

and the King of the cute stats, Aaron Schatz.

check this lil gem from last tuesday:
http://www.footballoutsiders.com/dvoa-r ... oa-ratings

"Atlanta is the worst 10-1 team DVOA has ever tracked, and it isn't even remotely close. Atlanta right now has 2.9% DVOA. The next-lowest team at 10-1 is the 1998 Vikings. Despite the record-setting offense, DVOA doesn't rate the Vikings that well, partly because the defense wasn't very good and partly because -- well, we're not really sure all the reasons why. But still, that Vikings team was at 17.0% after 12 weeks, sixth in DVOA. That's way ahead of the current Falcons. Atlanta actually has negative DVOA in its last four games, and even more surprising, six out of 11 games all season. So the team that currently has pole position for the number-one seed in the NFC has played only five above-average games all year according to FO stats."

(Cue Pudge's defense of DVOA, or DYAR, or whatever cute stats here) 8-)


two days later, the 'worst 10-1 team' beats the vaunted New Orleans Offense, destroying thier chances: and yesterday clinched the division with the Buccs loss to Denver. :roll:

That, kids is why you have to take 'cute stats' with a grain of salt. I am not discounting ALL of his points, the guy is somewhat fair. But, speaking of Denver, Mr Schatz gave Tim Tebow absolutely no chance to beat the Steelers in the playoffs due to his magical cute stats. Tebow goes in and destroys them, and this guy clams up. He pronounces players and teams 'finished' and then when he gets it wrong: you never hear him go 'oops'.

almost never. Take this hilarious journey into prognostications from August 1st:

http://www.thebiglead.com/index.php/201 ... -playoffs/

:rofl: :rofl: :rofl: :rofl: :rofl: :rofl:

cute stats! :rofl:

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 1:02 pm 
Offline
Purveyor of Truth & Justice
Purveyor of Truth & Justice
User avatar

Joined: Thu Dec 23, 2004 3:03 pm
Posts: 26068
Location: North Carolina
I'm going to defend it in this sense...

Your argument against cute stats apparently is that it is not perfect. Hmm, what is?
:?:

And when you actually read that Big Lead article, as opposed to just looking at the headlines, then you understand why some of his "bad" predictions were made. The Jets over the Giants? Well thats because of the belief that the Jets defense would be one of the best in the league. That of course was thrown out the window due to the injury to Darrelle Revis. Or do you think cute stats should be able to predict injuries as well?

As for the Bills, like most in the world, the perception again there was the addition of Mario Williams would make that one of the better defenses in the league, as the primary path to their greater success in 2012. That did not come to fruition, but again do you think cute stats should be able to predict things that the layman cannot?

You state that "cute stats" should be taken with a grain of salt. But I would say ALL football prognostications should be taken with a grain of salt, regardless of whether they are based off cute stats or the eyeball test or whatever.

Aaron Schatz gave the Broncos no chance of winning that playoff game last year. But guess what? So did 70% of the rest of America that weren't basing it off cute stats.

I'm waiting for your treatise on the alternative to cute stats that is so much better, and much more worthy of our attention?

And for the record, I think DVOA is still probably more accurate than traditional rankings/ratings such as Scoring Offense, Total Offense, Scoring Defense, and Total Defense.

_________________
"Vincere scis, Hannibal, victoria uti nescis" -- Maharbal, 216 B.C.E.


Top
 Profile E-mail  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 2:22 pm 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
Pudge wrote:
I'm going to defend it in this sense...

Your argument against cute stats apparently is that it is not perfect. Hmm, what is?
:?:

You state that "cute stats" should be taken with a grain of salt. But I would say ALL football prognostications should be taken with a grain of salt, regardless of whether they are based off cute stats or the eyeball test or whatever.


I'm waiting for your treatise on the alternative to cute stats that is so much better, and much more worthy of our attention?

.



:rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :rofl: :dance:

Image

a little touchy there, Pudge? :mrgreen:

First, I baited that thread and you took it hook, line and sinker! :clap:

I knew, absolutely knew, that you couldn't resist: and your response would be 'okay genius, where do you go for YOUR prognostications?'

Here is the problem: when you put yourself out there as some sort of mathematical 'expert' and claim your system is the 'best' then you have to take the heat when your 'system' gets burned. DVOA is okay, but much like new-school measurements PECOTA and CHONE for last baseball season, trusting supposedly air-tight applied math to win games doesn't always work out too well..They claim it's objective because it's all stats, but that's not exactly the case. The determination of how certain things qualify under their categories are very much subjective and opinion. The point values and determining a "successful play" to asses a plays value are quite subjective even though they try to claim otherwise.

So here is the part where Pudge pouts a little and demands to know who I would go to..My answer is simple: Vegas.

The legalized gambling circuit actually HAS $$ on the 'line' whereas guys like Profootballfocus and FootballOutsiders are actually getting paid to give thier picks. Rather then trust DVOA, or DYAR. or LGTB, or whatever: there are a number of Vegas sites that regularly perform better then any of those clowns. Sure, USA Today and ESPN cant use 'My Cousin Vinny' so they have to go to these guys. But sports book operators look at the entirity of the information ( yes Pudge, pat yourself on the back here because Vegas DOES read FootballClownsiders) but they also use thier 'eyeball test'. Vegaswatch is pretty good usually, and they have alot of complicated cuteness that would appeal to certain types, but I like to look to Jay Kornegay, who runs the sportbook at the Las Vegas Hilton, and Dave Tuley from ViewfromVegas. Dave appears regularly in ESPN, so he has access to some heady stuff. They take a look at each of the weekend's NFL games by getting the "public perception" from the consensus numbers at ESPN.com's Picks Central, which also has a lot of great information to help in your handicapping, and also gives the wiseguys' view of the game.

Pudge your right pobodys nerfect, and all prognostications should be taken with a grain of salt...If you look at this guy vs Football Outsiders, he regularly outperforms them..but some guys out there like to say they got a special system of looking at cute stats, and then never show up to take the fall..

:whistle:

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 5:31 pm 
Offline
Purveyor of Truth & Justice
Purveyor of Truth & Justice
User avatar

Joined: Thu Dec 23, 2004 3:03 pm
Posts: 26068
Location: North Carolina
Ok fine. Maybe DVOA is inferior when it comes to picking games. I wouldn't know, because I don't use them for that.

I don't really use any cute stats to determine that. I will use them to determine matchups, because that's how I pick games, because I know that the results of football games are determined by matchups, and nothing else.

So knowing that Tampa Bay is a run-first team, I might check out Football Outsiders to determine how good the Cowboys defense is against the run according to DVOA, to determine whether they can match up well against the Bucs run attack. I might also check out a number of other sites, notably Advanced NFL Stats to determine whether or not the Cowboys defense run success rate is good or bad. I would not use Pro Football Focus because their grades are determined by combining the sums of individuals, instead of measuring by units.

I'm not aware of Football Outsiders' claim that DVOA-based picking is "air tight" and thus my defense of DVOA is not built around that.

But I am aware of DVOA's claim that it is more accurate than traditional statistical measurement, such as total yards gained/allowed or total points scored/allowed in determining which teams are good/bad, and where unit-based strength/weakness lies (i.e. how well this team passes or defends the run), which is noted in its explanation. And I believe it is, because DVOA like most advanced or cute stats breaks down football on a play by play basis, rather than a game by game basis, which I think most people would agree is a much more accurate determination of a team's efficiency.

Again, while I'm not sure whether DVOA accurately projected the Broncos win over the Steelers, I'm fairly confident it did accurately do so the following week with the Patriots win over the Broncos.

So basically the idea of they got this notable pick/prediction wrong and therefore it negates the very existence of this metric, which appears to be your stance, is flimsy argument at best. Because didn't Vegas also heavily favor the Steelers in that playoff matchup last year? Weren't they favored by most sports books to win by more than a TD, like 8 points? Yes, I'm aware that the spread is different from picking the winner of a game. But that large a spread does often indicate that the majority of "experts" would pick the heavily favored team to win the game.

It's a tool to be used, and one that I think most would agree is a worthwhile tool. You could certainly sit back and claim that there are better tools. But your claim instead appears to be "this tool sucks" and I don't think you have any legs to stand on there. Or at least your argument thus far presented to support it doesn't have any legs.

I don't watch Numbers Never Lie on ESPN, where I know Aaron Schatz is a frequent guest (if not a co-creator). Maybe he and others out there are pimping DVOA to be the best tool out there. I'm sure because it's his baby, and Bill Barnwell of Grantland is a former FO-writer, they pimp it regularly in their writings/musings. But I don't regularly listen to that, because I know its far from perfect. Now maybe others aren't aware of this, and therefore need to be warned by good Samaritans like you fun gus of the "trappings of DVOA." There probably will never be a perfect metric, and if there will be we are currently a long way from it.

But in general, I'm a staunch defender of advanced metrics, because I believe they are superior to traditional stats that have been regularly a part of the national consciousness for decades. Relative to other sports, advanced metrics for football are in their infancy.

If your stance is "Beware of 'cute stats' because there are still flaws and kinks that need to be worked out" then I agree 100% with you.

But if your stance is "Beware of 'cute stats' because they are worthless" which is apparently what you were baiting me with your initial post, then I 100% disagree with you.

_________________
"Vincere scis, Hannibal, victoria uti nescis" -- Maharbal, 216 B.C.E.


Top
 Profile E-mail  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 7:21 pm 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
:lol: :clap: :dance: :mrgreen:

Image

You know Im just jerkin yer chain, right? Pudge: I am aware you dont like 'Domes' and yer not a fan of seeing us here in the playoffs. You have said as much, and I respect that. But I lived here when the Falcons played at FCS, and I dont think I went...I regret that.

The 'new dome' is as good as done. The 'powers that be' have determined it, and I have gone through all five stages. But: dont you want to see a playoff game in the GD before it's razed?

I wonder if you would consider a trip to the ATL vs a trip to N.O...Im going to both, either way.

now for your entertainment:


Image

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Mon Dec 03, 2012 11:23 pm 
Offline
Purveyor of Truth & Justice
Purveyor of Truth & Justice
User avatar

Joined: Thu Dec 23, 2004 3:03 pm
Posts: 26068
Location: North Carolina
I'm not a baseball nerd. But otherwise, guilty as charged. BTW, Emotions are for the weak!

As for going to a Falcon playoff game, if the offer is on table, I think I'll make my way down I-85 this year.

_________________
"Vincere scis, Hannibal, victoria uti nescis" -- Maharbal, 216 B.C.E.


Top
 Profile E-mail  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Tue Dec 04, 2012 12:17 am 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
Pudge wrote:
I'm not a baseball nerd. But otherwise, guilty as charged. BTW, Emotions are for the weak!

As for going to a Falcon playoff game, if the offer is on table, I think I'll make my way down I-85 this year.

:clap: :dance:

OK! sweet. I think were allowed to get some 'xtra' tickets, but don't be too excited: 300's,endzone corner.

I'd get em to you for face..As long as you deny ever liking DVOA..I kid, I kid

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Tue Dec 04, 2012 2:20 am 
Offline
Hall of Famer
Hall of Famer
User avatar

Joined: Sat Jan 29, 2005 7:12 pm
Posts: 6240
Location: Planet Claire
I thought cute stats involved the Dream Boat who, if I am not mistaken, is no longer with the team. If Pudge comes to ATL I need to press flesh with him. I've known him cyber space for half his life! :lol:

_________________
Image


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Tue Dec 04, 2012 8:25 pm 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
backnblack wrote:
I thought cute stats involved the Dream Boat who, if I am not mistaken, is no longer with the team. If Pudge comes to ATL I need to press flesh with him. I've known him cyber space for half his life! :lol:



hey, if I get 2 together would you two consider sittin next to each other? Cmon, it's be like Gandalf and Frodo!

BTW, Pudge we play this scenario out EVERY YEAR. Every year I have to ask how you think we will play in January, but what are your thoughts on exactly WHEN we would play. You know every damn Jan and Feb I have scheduling issues. But this year, if it's a home game playoff with a bye: Im staying and cancelling all work.

:dance:

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Tue Dec 04, 2012 8:44 pm 
Offline
Hall of Famer
Hall of Famer
User avatar

Joined: Sat Jan 29, 2005 7:12 pm
Posts: 6240
Location: Planet Claire
Image
I'd love to sit in between you guys but aguy at the Roost already offered to hook me up with two for every playoff game we have and I told my wife we'd go! She went with me when we beat the Vikes in 98 so her good luck is important for the franchise. But I must meet Pudge....that is if he doesn't quit watchign the game at halftime to go take a shower or go to a stat site! :wink:

_________________
Image


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Tue Dec 04, 2012 8:53 pm 
Offline
Draught Guru
Draught Guru
User avatar

Joined: Mon Oct 01, 2007 7:32 am
Posts: 5066
backnblack wrote:
Image
I'd love to sit in between you guys but aguy at the Roost already offered to hook me up with two for every playoff game we have and I told my wife we'd go! She went with me when we beat the Vikes in 98 so her good luck is important for the franchise. But I must meet Pudge....that is if he doesn't quit watchign the game at halftime to go take a shower or go to a stat site! :wink:



gotcha. after the Green Bay playoff game my wife swore off going with me...can't really blame her

:oops:

_________________
"what if there were no hypothetical situations?"


Top
 Profile  
 
 Post subject: Re: why 'cute stats' are misleading
PostPosted: Wed Dec 05, 2012 12:44 am 
Offline
Hall of Famer
Hall of Famer
User avatar

Joined: Sat Jan 29, 2005 7:12 pm
Posts: 6240
Location: Planet Claire
fun gus wrote:
backnblack wrote:
Image
I'd love to sit in between you guys but aguy at the Roost already offered to hook me up with two for every playoff game we have and I told my wife we'd go! She went with me when we beat the Vikes in 98 so her good luck is important for the franchise. But I must meet Pudge....that is if he doesn't quit watchign the game at halftime to go take a shower or go to a stat site! :wink:



gotcha. after the Green Bay playoff game my wife swore off going with me...can't really blame her

:oops:

:lol: My days of losing it at the Falcons probably ended with the 1900s.

_________________
Image


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 12 posts ] 

All times are UTC - 4 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  


cron